Tetrahydrocannabinol (THC), is known to have marked effects on discrimination of time duration. Six normal volunteers, (3 men) with no drug abuse history were administered oral Marinol (THC) in two doses (10mg or 20mg) or placebo, in identical masked capsules in a double blind manner prior to Oxygen-15 PET scanning. The study took place on three days, each separated by one week, with one placebo and two THC doses randomly assigned. All possible order combinations of the three doses were used across subjects. Two-and-a-half hours after oral drug administration, corresponding to maximal clinical effects, O-15 PET scans were performed. Three conditions pertained on each of the 3 study days: "baseline" (low sensory stimulation), passive tone-listening and tone-activated time estimation. Subjects were scanned twice in each non-baseline condition and scans averaged. Analog scales were used to quantify intoxication and THC blood levels taken at several time points. The computer program 'SPM 96' was used to determine regions of significant change in regional cerebral bloodflow, (rCBF). Pilot data analyzed on the first four subjects showed that high-dose THC compared to placebo in baseline condition, increased rCBF in cerebellum, orbito-frontal and posterior cingulate cortex (PCC) and decreased rCBF in the right fronto-parietal region. Regions showing increased rCBF during time estimation vs. passive tones on placebo included bilateral DLPFC and IPL. This condition was associated with reduced rCBF in bilateral fronto-temporal, Broca's area and PCC. These changes differed from those on high-dose THC, where time estimation (as defined above) produced increases in orbito-frontal and right cerebellum and left medial frontal gyrus, and decreases in right inferior fronto-temporal regions. Overall THC both impaired time estimation and altered rCBF patterns during time estimation. Regions of drug activation at baseline are consistent with known distribution of cannabinoid receptors, and known effects of marijuana of disrupting time estimation.